The gaps your tools cannot surface
Attributes with no engagement because there was never anything to engage with. They are invisible to click-ranked tooling and they are often the top repair.
How we compare
Tools that prioritise catalogue work rank by engagement, which is sensible and has one structural blind spot. It happens to be the expensive one, and it is the only reason to add us to a stack that already works.
An attribute recorded on zero products generates zero clickstream, so it scores zero and can never rise in an engagement-ranked list. The tool is working correctly; the signal is simply not there.
Blind to exactly the attributes worth adding.
Measured on the products themselves rather than on traffic, so it can value a field that has never been available to filter on, including one no shopper has ever had the chance to use.
Narrow, checkable, and the whole difference.
Where you are
Four things a catalogue has to do. Every tool here does three of them, and they are the same three, because all three learn from engagement that already happened.
If you run a PIM, a feed manager or a merchandising tool that ranks enrichment work, it is almost certainly ranking by engagement: filters used, clicks, channel performance. That is a reasonable way to prioritise and it works well for fields your shoppers can already reach.
The blind spot is structural rather than a matter of quality. A shopper cannot click a filter you never built. A field your catalogue has never recorded produces no engagement, so it generates no signal, so it never rises in a list ordered by signal. The fields most worth adding look least important by exactly the measure meant to find them.
That is the whole of our argument. We are not better at reading clicks. We start from the decision instead, which means a field recorded on zero products can still come out first.
On the shelf we measured, the top repair was missing from 80 of 216 products and had generated no engagement at all, because there was nothing to engage with.
The blind spot, in one line
Every tool in your stack ranks from demand it has already seen. That is a structural limit, not a gap in their effort.
80
Products missing it
of the 216 on the shelf we measured
0
Clicks it generated
there was nothing to click, so nothing was logged
#1
Its rank, ours
top repair, computed from the catalogue
One attribute, three ways of looking at it. An engagement-ranked tool sees the middle figure and concludes nobody wants it. These are three different units, so they are three tiles rather than one chart.
What it looks like
Missing on the most products. Needed by the fewest shoppers. The two rankings invert.
By engagement
By decision value
Missing on the most products. Needed by the fewest shoppers.
What happens
Engagement ranking handles these well. If a filter exists and people use it, click data tells you plenty and you do not need us for it.
Engagement under-weights these, because partial coverage produces partial signal. We score what completing them would be worth, which is usually much more.
Engagement cannot see these at all, by construction. This is where the ranking flips, and it is the only reason to run both.
What you get
Attributes with no engagement because there was never anything to engage with. They are invisible to click-ranked tooling and they are often the top repair.
Bits of decision information, so screen size and processor can be weighed against each other rather than both reported as a completeness percentage.
A rank with a price attached: what a script closes this week, against what has to be acquired. Most prioritisation tools order the work and leave you to cost it.
The output is a priority list. It goes into your PIM, your backlog or your merchandising plan. Nothing here asks you to move off anything.
Before you commit
Questions
One category export. If our order matches what you already knew, that is a useful answer too.
Request a Decision Audit